Disordered Function Conjunction: On the in-silico function annotation of intrinsically disordered regions.

Area of Science:

• Biochemistry
• Computational Biology
• Structural Biology

Background:

• Intrinsically disordered regions (IDRs) lack stable structures but perform vital biological roles, including molecular interactions and entropic functions.
• Current computational tools predict functions at the residue level, but annotating the vast number of discovered IDRs at the region level remains a challenge.

Purpose of the Study:

• To assess the feasibility of using residue-level prediction methods for predicting region-level functions of intrinsically disordered regions.
• To investigate the effectiveness of combining multiple residue-level predictions and sequence-derived features for multifunction prediction in IDRs.

Main Methods:

• Constructed a dataset of single-function IDRs, distinct from existing residue-level predictor training data.
• Evaluated the performance of available residue-level prediction methods for region-level function prediction.
• Incorporated amino acid composition alongside residue-level predictions to enhance classification accuracy.

Main Results:

• Residue-level prediction methods show modest utility for predicting multiple region-level functions of IDRs.
• Simultaneous use of multiple residue-level function predictions significantly enhances classification performance.
• Inclusion of amino acid content further improves the accuracy of multifunction predictions for IDRs.

Conclusions:

• Multifunction prediction for intrinsically disordered regions is achievable using current computational approaches.
• Leveraging existing residue-level predictors, combined with sequence features, offers a promising avenue for IDR functional annotation.
• Acknowledging and accommodating inaccuracies in functional annotations is crucial for reliable IDR multifunction prediction.